How a credit risk model can help investors improve stock selection

The StarMine Combined Credit Risk Model is a powerful tool to evaluate corporate credit risk. We discuss how this model can be extended to add value to equity portfolios.

1. The StarMine Combined Credit Risk Model (CCR) blends the strengths of the other three StarMine credit risk models to generate an estimate of public company credit risk.
2. Strong mean reversion is observed among low-credit-risk securities and enhanced momentum is observed among high-credit-risk securities in a combined value-momentum alpha model based top and the bottom decile portfolio.
3. The momentum effect is found to be stronger over one month while mean reversion is found to be stronger over 12 months.

For more data-driven insights in your Inbox, subscribe to the Refinitiv Perspectives weekly newsletter.

StarMine credit risk models

The StarMine suite includes four credit risk models:

• StarMine Structural Credit Risk Model (SCR) evaluates credit risk from the equity market’s view using a proprietary extension of the Merton structural default prediction framework that models a company’s equity as a call option on its assets.
• The StarMine SmartRatios Credit Risk Model (SRCR) uses financial ratio analysis for credit risk assessment and incorporates both reported information and forward-looking estimates using StarMine SmartEstimate.
• The StarMine Text Mining Credit Risk Model (TMCR) mines the language in textual data from multiple sources (Reuters News, StreetEvents, conference call transcripts, corporate filings, and select broker research reports) to evaluate companies’ potential financial distress.
• StarMine Combined Credit Risk Model (CCR) combines StarMine SCR, StarMine SRCR and StarMine TMCR to generate a single, final estimate of public company credit risk. It significantly outperforms the three individual credit models as well as alternative approaches such as Altman Z-Score-based estimates. Its superior performance through the COVID-19 pandemic is discussed in the white paper The Performance of Refinitiv StarMine Credit Risk Models.

StarMine Combined Credit Risk Model creates powerful default predictions and assessments of credit risk that are more accurate than using any one data source alone

Using StarMine CCR in equity selection

Although StarMine CCR was trained to predict company default events, it can also be used to complement a fundamental or quantitative equity selection strategy.

We use the StarMine Value Momentum (ValMo) model, which combines StarMine’s two valuation models (StarMine Intrinsic Valuation and StarMine Relative Valuation), along with StarMine’s two momentum models (StarMine Analyst Revision, and StarMine Price Momentum), into one powerful stock-selection model, as the representative alpha model.

As StarMine ValMo is already a combination model, it has proved to be a robust and steady performer in all regions.

To study the joint existence of the momentum and mean reversion effect in ValMo portfolios, we construct two low-credit-risk groups from within the ValMo score-based top and bottom deciles.

StarMine provides model scores as ranks between 1 and 100 with 1 representing a “bearish” and 100 representing a “bullish” score.

• LowValMo-HighCredit: This group is created by identifying low-default-risk securities (CCR score >=90) from within the ValMo score-based bottom decile (ValMo score <=10). This group represents securities that are negative-past-losers that also appear overvalued (low ValMo score) but have very low credit risk.
• HighValMo-HighCredit: This group is created by identifying low-default-risk securities (CCR score >=90) from within the ValMo score-based top decile (ValMo score >90). This group represents securities that are positive-past-winners that also appear undervalued (high ValMo score) and have very low credit risk.

The two groups are reconstituted monthly using the month-end date ValMo and CCR model scores.

Tables 1 and 2 below compare the forward-twelve-month return (f12M_ret) of the two groups with the remaining securities in their respective deciles (ExGrp) and equal-weighted market (Mkt) returns from January 2008 to September 2021.

We include the largest 3,000 U.S. equity securities by market capitalization as our analysis universe.

• Low-default-risk (high CCR score) securities are exhibiting strong mean reversion while ExGrp securities are exhibiting weak momentum over twelve months in both the top and the bottom ValMo deciles.
• The strength and the duration of the momentum and mean reversion is different among securities within the same decile. The mean reversion effect (in low-default-risk securities) is much stronger than the momentum effect (in ExGrp securities) over twelve months and seems to develop quicker in both the top and the bottom deciles.

Momentum and mean reversion effects are also observed over one month. However, unlike the twelve-month case, the momentum effect among high-default-risk securities is found to be stronger than the mean reversion among low-default-risk securities over one month.

Further, our research shows that the explicit inclusion of the mean reversion and momentum effects through CCR-ValMo interaction terms can quantitatively help in forecasting both longer-term (forward twelve months) and shorter-term (forward one month) holding period excess returns.

To learn more about the research, download our white paper: Using StarMine Combined Credit Risk Model in Equity Selection.